key: cord-0807299-8hk5snk2
authors: Krokhine, Sophie; Torabi, Hadis; Doostmohammadi, Ali; Rezai, Pouya
title: Conventional and Microfluidic Methods for Airborne Virus Isolation and Detection
date: 2021-07-02
journal: Colloids Surf B Biointerfaces
DOI: 10.1016/j.colsurfb.2021.111962
sha: 99f6701bb4b4241e06bb4c3167fde778222d3222
doc_id: 807299
cord_uid: 8hk5snk2

With the COVID-19 pandemic, the threat of infectious diseases to public health and safety has become much more apparent. Viral, bacterial and fungal diseases have led to the loss of millions of lives, especially in the developing world. Diseases caused by airborne viruses like SARS-CoV-2 are difficult to control, as they are easily transmissible and can circulate in the air for hours. To contain outbreaks of viruses such as SARS-CoV-2 and institute targeted precautions, it is important to detect them in air and understand how they infect their targets. Point-of-care (PoC) diagnostics and point-of-need (PoN) detection methods are necessary to rapidly test patients and environmental samples, so precautions can immediately be applied. Traditional benchtop detection methods such as ELISA, PCR and culture are not suitable for PoC and PoN monitoring, because they can take hours to days and require specialized equipment. Microfluidic devices can be made at low cost to perform such assays rapidly and at the PoN. They can also be integrated with air- and liquid-sampling technologies to capture and analyze airborne and body fluid-based viruses. Here, conventional and microfluidic virus detection methods are reviewed and compared. The use of air sampling devices to capture and concentrate viruses is discussed first, followed by a review of analysis methods such as immunoassays, RT-PCR and isothermal amplification in conventional and microfluidic platforms. This review provides an overview of the capabilities of microfluidics in virus handling and detection, which will be useful to infectious disease researchers, biomedical engineers, and public health agencies.

sampling and analysis methods. We discuss microfluidic applications of traditional techniques such as ELISA, PCR and LAMP as well as emerging techniques. We also discuss limitations and inconsistencies in previous works, such as the limit of detection reporting. At the end of this paper, we provide our assessments of future trends in the field of microfluidics for virus detection.

To monitor pathogens (including viruses) from the air and diagnose infections in patients, it is important to understand how they are spread. Viruses that spread through air, such as influenza and SARS-CoV-2, are the most easily transmissible. Transmission in air usually takes place in a four-step cycle ( Figure 1 ) involving 1) pathogen shedding by respiratory activities and formation of bioaerosols, 2) dispersion in the air, 3) infection of the recipient by inhalation or contact and 4) amplification of the pathogen within the host's body (Herfst et al., 2017) . J o u r n a l P r e -p r o o f First, pathogens are assimilated into bioaerosols in the respiratory tract, which are created by respiratory activities such as breathing, talking, sneezing and coughing (Morawska et al., 2009; Xie et al., 2009) . Bioaerosols are generally composed of one or more viruses along with mucus, salts and water and can range in size from several nanometers to several hundred micrometers (Holmgren et al., 2010; R. Lu et al., 2020) . After being released, they may either travel directly to the recipient or deposit on a surface and be re-aerosolized. The World Health Organization (WHO) classifies bioaerosols into two types, i.e., true aerosols (diameter of ≤5 µm) which can spread over distances greater than 1 m, and droplets (>5 µm) which can move over a shorter distance (Gralton et al., 2011) .

Aerosols tend to remain in the air longer than droplets and deposit in the lower respiratory tract, causing more severe clinical presentation (Gralton et al., 2011; Killingley and Nguyen-Van-Tam, 2013) . Different precautions are recommended for each transmission route; while surgical face masks and gloves may protect against droplet and contact transmission, N95 respirators may be needed to keep out aerosols (Ather et al., 2020 ; "N95 Respirators, Surgical Masks, and Face Masks | FDA," n.d.).

The 5 µm particle cutoff rule for aerosol classification does not take into account other factors such as viral load or environmental conditions which may influence deposition in the lungs.

Nevertheless, it provides a starting point for classifying bioaerosols based on size. The size and number of bioaerosols generated depends on the type and frequency of respiratory activities performed (Duguid, 1946; Morawska et al., 2009; Xie et al., 2009) . For example, normal breathing usually produces aerosols from the upper respiratory tract, while talking, sneezing or coughing generates larger droplets from the movements of the vocal cords (Cowling et al., 2013; Herfst et al., 2017) Aerosols can also be generated by healthcare procedures involving the lungs, such as ventilation, intubation and bronchoscopy (Ather et al., 2020) . The influence of different respiratory J o u r n a l P r e -p r o o f activities and environmental conditions on particle composition is reviewed in detail elsewhere (Gralton et al., 2011) . Individual differences such as age also contribute to the variability in the number of particles exhaled, and "super-spreaders" (people who may exhale disproportionate amounts of particles and infect many others) contribute in large part to disease outbreaks Small et al., 2006) .

After generation, the bioaerosols are dispersed in the air at low density. Bioaerosols usually absorb water from the mouth or respiratory tract, in a process known as hygroscopic growth (Morawska et al., 2009) . Once they are released, the water may evaporate, thus reducing their size (Gralton et al., 2011) . Relative humidity affects hygroscopic growth and evaporation, such that aerosols remain larger at higher humidity (Gralton et al., 2011; Morawska et al., 2009 ). Viruses survive optimally at different temperatures and humidity levels. Enveloped viruses such as influenza and coronaviruses typically survive longer at lower temperature and humidity, while non-enveloped viruses such as rhinoviruses are more stable at higher humidity (Casanova et al., 2010; Niazi et al., 2021) . At low relative humidity, the salts inside aerosols and droplets may crystallize, which can protect viruses from degradation. The type of ventilation and air conditioning used indoors can also enable viruses to spread further and survive for longer, increasing the risk of transmission (Ather et al., 2020) . The effect of environmental factors on particle size is reviewed more comprehensively elsewhere (Gralton et al., 2011) .

Next, the pathogen deposits in the recipient, usually via inhalation or contact, and infects the respiratory tract. Bioaerosols can be inhaled directly from air containing them; they also often deposit on surfaces such as walls and objects including cellphones and door handles, as has been seen with COVID-19 (Lane et al., 2020; Mouchtouri et al., 2020; Razzini et al., 2020) . Once a surface is touched, the virus can be re-aerosolized or introduced back into the air. A person's chance J o u r n a l P r e -p r o o f of becoming infected depends on the number of pathogens they are exposed to. The 50% human infectious dose (HID50) is the dose corresponding to a 50% chance of infection in a susceptible individual (Yezli and Otter, 2011) . HID50 is different for different viruses; for example, the HID50 for influenza virus is much greater than that for rhinovirus (Yezli and Otter, 2011) .

Finally, viruses are amplified or replicated in their new host, in either the respiratory tract or some secondary area. They can "hijack" cells' genetic machinery to replicate themselves through the lytic or lysogenic cycle (Lodish et al., 2000; Louten, 2016) . In most cases, they will infect many of the cells in a given area, which sometimes but not always leads to symptoms. Viruses can also infiltrate immune cells to protect themselves from attack. Once viruses have been significantly amplified, they are shed through bioaerosols in the upper respiratory tract. Asymptomatic people can still shed the virus without knowing they have it, and can transmit it to their close contacts (Cowling et al., 2013; Ji et al., 2020; Kenarkoohi et al., 2020; Nishiura et al., 2020) . This is especially important for the spread of COVID-19.

Understanding the transmission cycle and specific modes for each pathogen is important to plan appropriate precautions. Air monitoring focuses on identifying viruses at the second step, after they have been shed but before they have been deposited in the recipient. Disinfection, remediation and evacuation or quarantine measures can be applied to stop the transmission cycle, and prevent many infections. Various methods for virus collection and detection in indoor and outdoor air samples are described in the following sections.

The majority of methods for virus analysis and detection require samples to be in a liquid or swab form, so aerosol and droplet particles must be collected and concentrated in a separate step.

J o u r n a l P r e -p r o o f Several types of commercially available samplers can be used to do so, including impactors, cyclones, impingers, filters and electrostatic precipitators ( Figure 2 and Table 1 ) (Pan et al., 2019; Verreault et al., 2008) . Selecting samplers for PoN testing or research requires understanding of their collection efficiency (J. . Two types of collection efficiency are measured, i.e., physical (the ratio of particles in the environment to particles captured by the sampler) and biological (the percentage of virus that stays viable after collection). Physical collection efficiency is usually measured by counting the numbers of particles entering and exiting the sampler, while biological collection efficiency is often measured by culture or plaque assay. 

The most commonly used sampling devices are liquid impingers, which work by forcing air containing viruses through nozzles into a collection medium ( Figure 2 ) (Mirzaee et al., 2016; Pan et Cascade Impactor Impinger Electrostatic Precipitator

Water-based growth tube collector J o u r n a l P r e -p r o o f al., 2019). The air flow rate differs from device to device (Pan et al., 2019; Springorum et al., 2011) .

The pressurized air forms bubbles on the surface of the collection medium, which can allow small particles to diffuse into the medium (Verreault et al., 2008) . Most impingers are made of glass, although some can be made of metal (Pan et al., 2019; Verreault et al., 2008) . The SKC BioSampler is a commercial glass impinger, used as a reference sampler in many of the studies described in this review ("BioSampler," n.d.; Pardon et al., 2015) . It is designed to trap larger particles through impaction like a human respiratory tract, and it operates at a recommended flow rate of 12.5 L/min ("BioSampler," n.d.; Verreault et al., 2008) . Impingers generally preserve viral infectivity, and their collection media can be directly harvested to perform an assay Riemenschneider et al., 2010) . However, wall loss (adherence of virus to sampler wall preventing capture into the media), and re-aerosolization can be problematic, leading to lack of detection of positive samples and underestimation of the true concentration of viruses (Grinshpun et al., 1997; Hong et al., 2015; Springorum et al., 2011) . Evaporation of the impinger liquid may also take place at higher temperatures.

Impingers are often used to detect viruses such as SARS-CoV-2 because the liquid medium preserves viability (Springorum et al., 2011 ). Faridi et al. (2020 used standard midget impingers at a flow rate of 1.5 L/min for sampling SARS-CoV-2 from 28 air samples from hospital wards and reported no virus detection using RT-PCR (discussed later). Ma et al. (2020) used automated impingers to collect 26 air samples, along with surface swabs and exhaled breath condensates, in hospitals and COVID-19 isolation centers. Samples were analyzed with PCR, and a lower positive rate was found for air samples (3.8%) than swabs (5.4%) and exhaled breath condensates (26.9%), possibly due to ventilation or inactivation by disinfectants. These results indicate that viruses were J o u r n a l P r e -p r o o f exhaled in a short time period, and further affirm the importance of bioaerosol transmission in the spread of COVID-19. Mirzaee et al. (2016) developed an impinger integrated directly on a microfluidic chip, which required significantly less time and reagent volumes than conventional impingers. The PDMS chip included intersecting gas and liquid channels, as well as gas and liquid inlets and outlets. Air containing 0.5, 1 and 2µm polystyrene latex (PSL) particles was drawn into the chip by a vacuum pump operating at flow rates of 10-20 mL/min. Controlled bubbling at the ends of gas channels was used to trap small particles. Both mathematical modeling and experimental results indicated a physical collection efficiency around 90%.

Impactors also involve air being drawn through a vacuum pump and pushed through nozzles, but onto solid surfaces rather than liquid collection media ( Figure 2 ) (Hong et al., 2015) . The deposition of particles onto the surface is dependent on their inertia and mass, such that only particles within a certain size range may impact onto a specific surface (Herfst et al., 2017) . The Andersen

Cascade Impactor (ACI) separates particles by size using a series of six surfaces organized serially in the air flow direction; the largest particles impact on the first surface and particles of successively decreasing size deposit on the lower surfaces (Andersen, 1958; Park et al., 2016) . Unfortunately, some small virus-containing particles cannot impact on even the last surface and must be collected after by other methods like filters (Appert et al., 2012) . Similar to impingers, impactors suffer the drawback of significant wall loss. Hong et al. (2015) developed an inertial size separator directly on a microfluidic chip. The chip had a main curved channel with 3 outlets designed to separate bacteria (S. epidermis) from viruses (Adenovirus 40; Figure 3 ). The outlets were positioned at 90° curves, and aerosols were J o u r n a l P r e -p r o o f 13 pumped first at 120 mL/min to separate large particles (>3μm) at the first outlet, then at 160 mL min -1 to separate bacteria-sized particles (1-2µm) at the second outlet. Virus-size particles (<500nm) exited at the third outlet. Gelatin filters were placed at each outlet to trap exiting particles for qPCR analysis. Results were acceptable with approximately 70% of 3.25µm PSL particles exiting at the first outlet, 78% of bacteria exiting at the second outlet and 68% of viruses exiting at the third outlet.

Bioaerosol losses were relatively low, i.e., 3.8% for viruses and 3.5% for bacteria. qPCR results confirmed that bacteria and viruses were efficiently separated. Cyclones, also known as centrifugal samplers, are circular samplers that operate similar to impactors ( Figure 2 ) (Cho et al., 2019; Lane et al., 2020) . Centrifugal force disrupts the flow of air containing particles in cyclones, causing the particles to impact upon the collection wall. Cyclones are not very efficient for collecting low concentrations of virus particles; this may be why Faridi et al. (2020) found no positive SARS-CoV-2 samples in a patient room using 2-stage NIOSH cyclones.

Traditional cyclones may desiccate viruses thus decreasing their infectivity, but "wet cyclones", in J o u r n a l P r e -p r o o f which particles impact onto liquid media such as water, may provide gentler collection (Cho et al., 2019; Orsini et al., 2008) . Cho et al. (2019) created a device called the Automated and Real-Time Bioaerosol Sampler based on Wet-Cyclone (ARBSW) that was integrated with a microfluidic flow cytometer. The ARBSW consisted of a plastic funnel-shaped apparatus, with water forming a liquid film on the surface. The stark difference between the flow rates of the air, 16 L/min, and the liquid, 9 mL/hr, caused particles to quickly impact on the film and transfer to the flow cytometer. S. epidermis and M. luteus captured by the sampler were cultured, and the number of colony-forming units (CFUs) were counted. Sampling was carried out within 20min, and the physical collection efficiency for bacteria was ~95%. Microbial recovery was similar to that of the reference BioSampler, suggesting that this technology should be explored more in the future.

It is often difficult to collect small particles with impingers or impactors, so filters provide an alternative (Kwon et al., 2014; J. Li et al., 2018) . Filters can be made of many different materials such as cellulose, glass fiber, mixed cellulose ester (MCE) and polytetrafluoroethylene (PTFE; Figure 2 ). The mechanisms by which different types of filters remove particles are reviewed elsewhere (Verreault et al., 2008) . Although convenient, they tend to dry out viruses and the process of extracting viruses can both damage their infectivity and eliminate the possibility of a direct sampling-to-analysis workflow. Gelatin filters can be dissolved in liquid post-capture and preserve viability; therefore, they are the preferred filter type for studies investigating infectivity (Mouchtouri et al., 2020) . Gelatin filters inside an MD8 Sartorius sampler have been used to detect SARS-CoV-2 from hospital air (Mouchtouri et al., 2020; Razzini et al., 2020) . Mouchtouri et al. (2020) tested 12 J o u r n a l P r e -p r o o f air samples in a hospital setting. One contained viral RNA; it was captured 2.5m away from a patient not wearing a mask. Razzini et al. (2020) also detected positive air samples in the ICU and corridor of a hospital. Nevertheless, because of their propensity to dry out or melt, they should be used for only short periods of time at moderate temperature and relative humidity (Appert et al., 2012; J. Li et al., 2018) .

Another type of sampler that has recently become popular is the electrostatic precipitator (ESP), which uses electrostatic attraction to transport particles to a collection electrode (Figures 2 and 4) (Foat et al., 2016; Sandström et al., 2008) . Metal needles near the inlet of the ESP create a corona discharge which imparts charges on aerosols, so that they move towards the oppositely charged electrodes. Since ESPs carry less risk of damage to viruses than impactors or filters, create more concentrated samples, and are commercially available, they have commonly been integrated with microfluidic detectors (Pardon et al., 2015; Tan et al., 2011) . However, the movement of particles by electric field may cause damage to viral surface proteins and reduce infectivity, making certain assays more difficult to perform (Bhardwaj et al., 2020) . Figure 4A ). Three corona discharge needles with an inter-electrode distance of 3cm, as well as a liquid collector electrode, were used. The cylindrical sampler had two inlets to draw in particle and sheath flow, and one outlet was connected to an SKC BioSampler impinger. An aerosolized dye was used for testing, and particles that were not collected by the ESP were captured by the BioSampler. When the corona system was on, the ESP sampler demonstrated a collection efficiency of more than 20%. When the system was off, the collection efficiency remained below 1%, confirming that the use of the corona discharge greatly increased particle capturing.

J o u r n a l P r e -p r o o f Tan et al. (2011) designed an automatic electrostatic sampler (AES) for collection of pathogens from the air The AES consisted of a semi-spherical steel electrode containing a central copper plate, with diameters of 6mm or 16mm. Air was brought into the sampler at flow rates of 1.2 or 6.2 L/min and passed through a particle charger consisting of two copper needles. Non-biological particles of 0.3-20µm were used for testing the physical collection efficiency, and bacteria were sampled inside and outside to test the biological collection efficiency. For the 16mm central electrode when particle charger was applied, physical collection efficiency was higher at 1.2 L/min (above 90%) than at 6.2 L/min (60%). Fewer viable bacteria were collected indoors and outdoors by the AES than MCE filter, for both flow rates. The authors suggested that the AES could be integrated directly with an immunosensor for analysis. The 3D printed battery-powered samplers was operated at a flow rate of 5 L/min; air was drawn in through a small fan and passed through an inlet with corona needles. Bacillus atrophaeus spores, Pseudonomas bacteriophage 6 and sodium fluorescein were used in testing. A digital microfluidics (DMF) system known as electrowetting-on-diode (EWOD) was used to transport the sampled particles for analysis. In EWOD, particle-containing liquid from the sampler was formed into 2-3µL droplets which were actuated across an electrode surface. These droplets were then used to perform reactions and assays. Collection efficiency for sodium fluorescein aerosols reached about 80% but only for particles larger than 4µm. Biological collection efficiency was quite low for Bacillus atrophaeus spores (2.7%). Follow up tests determined this may have been due to corona discharge damaging the hydrophobicity of the actuation surface, which should be solvable.

In the study by Park et al. (2016) a single-stage electrostatic precipitator, called an "aerosolto-hydrosol" sampler, was used to capture airborne S. epidermis ( Figure 4B ). The aerosol-to-J o u r n a l P r e -p r o o f hydrosol sampler was composed of a polycarbonate sheath with a liquid sampling well inside, and it operated at a flow rate of 8 L/min with an applied voltage of -7 kV. Aerosols passed through the sampling well into a stainless-steel ground electrode, where they were captured in a flow of liquid containing lysis buffer and fluorescent reagents.

Various other sampling devices have been used for airborne viruses. One example is the water-based growth tube collector; in which aerosols are placed in a tube surrounded by cooled water Hering et al., 2014 . Water vapor in the tube condenses, capturing the particles in larger droplets which impact onto the wall of the tube. This can be used to quantify bioaerosols in condensation particle counters (CPCs) . Some researchers have created "custom samplers" that combine several of the above techniques (Damit, 2017; Novosselov et al., 2014) . Novosselov et al. (2014) used a W-shaped microchannel collector (µCC), in which particles impacted onto the wall by centrifugal force. The collection efficiency was about 50% for 0.5µm PSL particles and close to 100% for 2µm particles, and the authors suggested that such a collector could be integrated into microfluidic systems. Damit (2017) successfully distinguished aerosols containing dead E. coli from non-biological aerosols using a droplet-based microfluidic platform. Droplets containing the fluorescence agent propidium iodide were created at the intersection of fluid and oil channels. Both types of aerosols were sprayed from above and absorbed onto the surface of droplets. The droplets containing E. coli were selectively stained and had a bulk fluorescence measurement about 20-30 times greater. 

Once viruses have been captured in sampling, there are multiple commonly used methods to detect and analyze them. Such methods include immunoassays, involving detection of specific antigens or antibodies, nucleic acid amplification, involving copying and detection of specific regions of a viral genome, and visualization via microscopy. In this section, we will describe these methods in the context of conventional assays as well as their implementation or integration into microfluidic devices. These methods, and their applications to microfluidics, are also summarized in Table 2 .

Many biosensors have used the natural properties of the immune system, involving antigenantibody interactions, to detect and quantify pathogens. Examples of this technique include enzymelinked immunosorbent assay (ELISA), surface plasmon resonance (SPR) detection and lateral flow immunoassays that are described below.

J o u r n a l P r e -p r o o f ELISA is commonly used for detecting pathogens including viruses. It operates on the principle of specific antigen-antibody binding, i.e. antibodies bind to a certain region (epitope) of an antigen, and each antibody only recognizes one or a group of antigens (Hnasko et al., 2015) .

Conventional (benchtop) ELISA can take several hours, too long to quickly catch the spread of disease (Y. .

Microfluidic devices have been used to perform ELISA in real-time, as in the study by Dimov et al. (2020) . A DMF platform based on EWOD was introduced to perform automated ELISA on four different targets; although the samples were not airborne, one of the targets was the MS2 bacteriophage. ELISA was performed using MS2-specific antibodies immobilized on magnetic beads. Minimal volumes of reagents were used, and the entire assay could be completed in 6 to 10 minutes. Y. used reciprocating-flow ELISA on a microfluidic chip for detection of SARS-CoV-2 antibodies from the serum of 13 patients within 5 minutes.

Pressure was exerted and removed from the fluid, allowing it to flow back and forth and improving binding to immobilized antigens. The device achieved 100% sensitivity and specificity and a LOD of 4.14 pg/mL. In another microfluidic device developed by Yanagisawa and Dutta (2011) , kinetic ELISA was used to detect blue tongue virus (BTV) and epizootic hemorrhagic disease virus (EHDV) antibodies, which were extracted from the body fluids of mice and rabbits.

Capture antibodies were first immobilized onto a glass microchip with etched channels, and the chip was then incubated in solutions of BTV and EHDV antibodies. In kinetic ELISA, fluorescence was measured 6-7 times during the reaction period, nearly in real-time. It linearly increased with time,

showing the progression of the reaction. Kinetic ELISA was also performed on a commercial microwell plate for comparison. The microfluidic platform showed significantly better performance than the microwell, with a 3x lower limit of detection (LOD).

While ELISA is commercially available, even on a microfluidic chip it is not always suited for rapid PoC and PoN diagnosis. Before performing ELISA, researchers must have antibody solutions (antisera) available (Hnasko et al., 2015) . The long and costly process of creating antisera involves immunizing animals, often mice, against a pathogen multiple times before isolating antibodies from the animals' blood (Boonham et al., 2014; Hnasko et al., 2015) . Additionally, ELISA may not work well for detecting all types of viruses (Boonham et al., 2014) .Overall, many techniques were found to be useful for detection of viruses from air (Bhardwaj et al., 2021; Nath et al., 2020; Zhuang et al., 2020) . However, effectiveness of various devices is difficult to compare because the sensitivity, specificity and LOD are measured differently in these studies. LOD is measured in different units: e.g., copies per reaction, copies per mL, pg per mL, PFU, and CFU. Additionally, cutoffs for positive results and distinction of virus presence from background signals differ across studies. The use of a standard unit of LOD and signal threshold would make it easier to report results and detect viruses.

Immunoagglutination, which involves the detection of clumps of magnetic beads conjugated to antibodies, is a technique used to carry out immunoassays such as ELISA which has been used for detection of viruses including influenza (Kwon et al., 2014; P. H. Lu et al., 2020) . It is often used in combination with a DMF platform as in the studies by Coarsey et al. (2019) While ELISA is commercially available, even on a microfluidic chip it is not always suited for rapid PoC and PoN diagnosis. Before performing ELISA, researchers must have antibody J o u r n a l P r e -p r o o f solutions (antisera) available (Hnasko et al., 2015) . The long and costly process of creating antisera involves immunizing animals, often mice, against a pathogen multiple times before isolating antibodies from the animals' blood (Boonham et al., 2014; Hnasko et al., 2015) . Additionally, ELISA may not work well for detecting all types of viruses (Boonham et al., 2014) .

The SPR immunosensor was first commercialized in the 1990s as an alternative to ELISA (Schasfoort, 2017) . In an SPR apparatus, light is beamed at a sensor covered by metal film at a specific angle, and the intensity of light being reflected is measured. Of interest is the angle at which surface plasmons are excited and the lowest refraction intensity is achieved, known as the resonance angle or "SPR-dip". The resonance angle is affected by the refractive indices of metal on both sides of the sensor. When introduced antigens bind to ligands on the sensor surface, they change the refractive index on one side, causing a change in the resonance angle and enabling detection of the pathogen. A regeneration solution is then added to wash off bound antigens and allow the sensor to be reused, although ligands or metal coating may also be washed off in this step reducing the efficacy of the sensor. Detection is label free, meaning that no enzymes or other conjugates are needed (Chang et al., 2018) . Usachev et al. (2013) used a microfluidic SPR sensor coupled with an air sampler for realtime detection of MS2 aerosols. Once aerosols were generated by a nebulizer, they were captured by a 'bubbler' air sampler similar to that used by Agranovski et al.(2002) . Sampling times were 1, 5 and 25 min and the flow rate used was 4 L/min. MS2 could be qualitatively detected after 2 min, and the entire detection and analysis process was completed within 6 min. The sensor was specific to MS2, exhibiting no cross-reactivity with influenza A viruses or m13 and T4 phages. It was also shown to be durable, exhibiting a relatively small decrease in performance of 30% Many variations of immunoassays that do not fit into the above categories have been used.

Several studies using these non-standard immunoassays are reviewed below.

Chemiluminescence assays (CLIs) involve a chemical reaction between antibodies or antibody fragments and specific labels, generating luminescence. Labels include luminol, actinidium ester and metal-conjugated magnetic particles, and enzymes may catalyze the light-producing reactions (Nicol et al., 2020) . antibodies, which attached to the H1N1 particles. Virus-containing 5μm droplets were generated by a nebulizer and captured by a button filter operating at 10 L min -1 . Mie scattering from the beads was measured by both a miniature spectrometer and a cellphone camera, then was used to quantify the virus concentration. It was able to detect viruses in the presence of dust particles and worked optimally at a scatter angle of 45°. Very low LODs of 4 pg/mL with the miniature spectrometer and 10 pg/mL with the cellphone camera were reported. Ventilation, which is known to potentially cause the spread of viruses, was performed at both minimum (4.25 L/min) and high (29.7 L/min) rates. results in about 10 minutes, and are easy to manufacture and mass-produce. Because they come in the form of small strips, they are also portable and can be distributed to low-resource locations. In SARS-CoV-2 detection, antibody tests by LFIA are used to identify those who have been previously infected or to diagnose those who are negative by RT-PCR (Nicol et al., 2020) . IgG, IgM and/or IgA are usually detected from whole blood, serum or plasma within 10-15 minutes.. Nicol et al. (2020) used ELISA, CLIA and LFIA to detect SARS-Cov2 IgM, IgA and IgG antibodies. Sensitivity for all three methods and all three antibody types was below 60% in the first 7 days after symptom onset and reached 100% 14 days after symptom onset. Specificity for IgG was lower for ELISA (96.7%) than CLIA (99.3%) and LFIA (98%). In another study, the sensitivity to SARS-CoV-2 IgG antibodies reached 100% after 21 days . Perfect specificity was observed in all time intervals after symptom onset. LFIAs have also been used to detect other human viruses such as Hepatitis C and Zika from serum, with high specificity and sensitivity, and they could be used to analyze concentrated samples collected from air Xiang et al., 2012) . Xiang et al. 

In addition to antibody-and antigen-based assays, nucleic acid amplification is used to detect pathogens. PCR has been one of the most commonly used lab techniques since its inception in the 1980s as reviewed before (Singh et al., 2014) . PCR can create millions of copies of target DNA or RNA regions for later quantification with gel electrophoresis.

Conventional PCR is a long process that can take hours to complete, but quantitative PCR (qPCR)

can detect products in real time and significantly speed up diagnosis (Huijskens et al., 2012; Shen et al., 2011) . Currently, qPCR is the 'gold standard' for COVID-19 diagnosis, but the risk of healthcare worker infection from collecting samples has necessitated alternative diagnostic methods (Faridi et al., 2020; Farshidfar and Hamedani, 2020; Lane et al., 2020; Ma et al., 2020; Mouchtouri et al., 2020; Razzini et al., 2020) . It has also been used to detect various human viruses including rhinovirus, respiratory syncytial virus and cytomegalovirus (Huijskens et al., 2012; . For example, Huijskens et al. (2012) used PCR to detect 19 respiratory viruses and M.

pneumoniae in the nasal secretions of 177 children, in a retrospective study. 73% of children had at least one virus, mainly RSV (37%) and Human Rhinovirus (24%). Those with a respiratory pathogen were more likely to be hospitalized and present with symptoms like rhinorrhea and dyspnea.

Conventional PCR is not sufficient for rapid diagnosis in resource-poor areas because it is expensive and requires a skilled user and a centralized laboratory (Singh et al., 2014) . However, performing PCR on microfluidic chips may provide a way to overcome these limitations. Foat et al. immunofluorescence, RT-PCR, and virus isolation (to confirm viability). The chip was also integrated with next-generation sequencing (NGS) to identify a novel virus in turkeys. Once the virus was captured and sequenced, it was found to be a type of infectious bursal disease virus.

J o u r n a l P r e -p r o o f

Isothermal amplification, in which nucleic acid segments are amplified at a constant temperature, is an alternative to PCR which does not require the same cost and specialized equipment. The most commonly used amplification technique is LAMP, which is discussed below.

In LAMP, forward and backward primers are used to create looped segments of DNA which can then be detected by fluorescence or other methods (Notomi et al., 2000) . LAMP is conducted at a constant temperature of 60-65°C and generally takes 100 min or less to complete (Coelho et al., 2017; Fujino et al., 2005; Liu et al., 2016 Liu et al., , 2020 Sun et al., 2020; Wang et al., 2018) . LAMP has been widely integrated with microfluidic chips for pathogen detection due to its low LOD, low cost, and ease of use. It has been performed on microfluidic chips for SARS-CoV-2 detection, with test results generated in 20-70min ( Pseudonomas aeruginosa bacteria. A portable system was constructed containing a microfluidic chip along with a central circuit board, heater and detection modules. Air containing bacteria was vacuum-pumped onto the chip, passing through enrichment channels before reaching the LAMP detection chamber. LAMP products were illuminated by a fluorescent dye, SYBR green 1 (Zipper et al., 2004) , to enable detection. Results were viewed on an LCD screen integrated into a box, and the bacteria was reliably detected within 70min. In a later study, the same authors detected

Staphylococcus aureus, Klebsiella pneumonia, Acinetobacter baumanii and P. aeruginosa from the air using disposable microfluidic chips (Q. Liu et al., 2018a) . Detection was semi-quantitative, and standard quantity-time curves were generated for each species of bacteria. LODs of 50 or fewer copies in 6.6μL reaction volume were observed for each species.

In a second study (Q. Liu et al., 2018b) , a semi-porous membrane (microfilter) was integrated onto microfluidic chips for pathogen concentration and enrichment. Airborne P. aeruginosa was detected as in the previous study, with an observed collection efficiency above 99%.

J o u r n a l P r e -p r o o f 

Another technique for detection of viruses, and other nanoparticles and micro-organisms, is to visualize them directly using microscopy. Many microscopes, such as electron microscopes, are sensitive enough to allow viewers to see nanoparticles and viruses (van Helvoort and Sankaran, 2019). Because many of these sensitive microscopes are expensive and require a trained user, lensfree holographic microscopy and smartphone attachments have also been developed.

Holographic microscopy works by illuminating a sample (which can be liquid, in a disposable microfluidic device) from below and then digitally reconstructing the diffraction pattern into an image, or hologram. These microscopes are portable and often do not include lenses, but nanolenses may form when a Lamb-type wave reshapes the fluid on the substrate, making particles easier to view as shown in Figure 6 (Ray et al., , 2017 . Ray et al. (2017) used a holographic microscope J o u r n a l P r e -p r o o f for counting and estimating the concentration of herpes simplex virus, in which virus particles were conjugated to magnetic beads and then attached to antibodies on a glass surface. The recovery rate was lower than expected due to possible virus disintegration, difficulty with separating viruses which were close together and background signal in the field of view.

Microscopy can also be performed with apparatuses attached to a smartphone, to reduce cost and increase portability. In the study by Wei et al., (2013) , a smartphone-integrated apparatus was used for detection of fluorescently labeled human cytomegalovirus (HCMV) particles. It consisted of a laser excitation diode, an interference-based filter, an external lens, and a focusing stage.

Aggregated virus particles were associated with greater fluorescence intensity than single ones, and the density of viral particles measured with the microscope was strongly correlated with initial viral concentration. Ming et al., (2015) used a quantum barcoding system for multiplex detection of various viruses. Polystyrene nanoparticles were infused with quantum dots to act as barcodes for each virus, and had specific probe molecules which recognized and bound to viral genetic material.

The barcoded beads created different colors of fluorescence after laser excitation. Multiple emission filters were used to reduce background noise and distinguish viruses. Images were taken by an iPhone camera, and an analysis algorithm was used to determine the presence of viruses. HIV, HBV, HCV, influenza A H1N1, H5N2 and H3N1 and influenza B viruses were successfully detected, and HIV and HBV clinical samples were detected with 100% sensitivity and specificity and a LOD below 1000 copies/mL While these studies only analyzed virus samples from liquid, we believe that both types of microscopy could be used to analyze the liquid collected from air samplers.

J o u r n a l P r e -p r o o f 

Over the past few years, research in virus detection and diagnosis has expanded and advanced biosensing techniques have risen to more widespread use. Such techniques can be used to replace, or to complement, more conventional techniques discussed earlier in this review.

Surface-enhanced Raman spectroscopy (SERS) is the use of a special substrate with Raman spectroscopy to greatly amplify the Raman signal via the SPR effect. It can either be used on its own or to enhance sensitivity of LFIA. SERS has been used to create aptamer-based sensors for influenza virus Kukushkin et al., 2019) . In the study by Kukushkin et Field effect transistor (FET) based sensing is also a promising method for virus detection and analysis (Cui et al., 2001; Shen et al., 2012 Shen et al., , 2011 Uhm et al., 2019) . In FET, surfaces are modified by the addition of receptors or ligands to which target analytes bind, inducing a conductance change which is measured. FET can be done with carbon nanotubes, silicon nanowires (SiNW) or graphene as reviewed in (Poghossian et al., 2020) . Seo et al. (2020a Seo et al. ( , 2020b 

Microfluidic devices are well suited to monitor the spread of infectious viruses in the air and provide rapid diagnosis; the importance of which has been highlighted during the COVID-19

pandemic. These devices are portable, low-cost and can carry out reactions very quickly. Many advances have been made in sampling viruses directly from the air and concentrating these samples for downstream analysis. On-chip, a wide variety of analysis techniques such as immunoassays, RT-PCR and LAMP have been used which have enabled detection within minutes. This has included novel COVID-19 tests; however, few commercially available methods for SARS-CoV-2 detection from air currently exist.

Usually, air sampling is done before analysis. Therefore, this review focused on both sampling and macro to microfluidic analysis techniques. We reviewed several sampler types including the liquid impinger, impactor, cyclone and electrostatic precipitator. In some cases, the sampler was integrated directly with a microfluidic device, although this was rare. While samplers could generally produce concentrated samples for direct analysis, most damaged or inactivated viruses, which could render them less useful for studies of infectivity. Additionally, no sampler was found that worked well over the whole size range of particles (about 10 nm to several μm).

We then moved to reviewing bioanalysis techniques such as PCR, ELISA, LAMP, SPR and holographic microscopy. Many of these techniques have on-chip and off-chip (benchtop) versions.

While many of these studies did not use air samples, they were still reviewed here as the techniques J o u r n a l P r e -p r o o f could be useful for analysis in air. These tests can effectively detect viruses within minutes, and applications for SARS-CoV-2 detection were highlighted.

Overall, air samplers and microfluidic devices have proven useful for virus detection. They can help curb virus outbreaks, as is being seen with novel tests for COVID-19. These technologies will continue to be developed and used for years to come.

Microfluidic technologies need to be further adapted for combatting COVID-19 and for use in the post-pandemic period. Specifically, rapid air monitoring and PoC diagnosis for known and emerging viruses will be needed to detect and control future outbreaks in their early stages. Devices should be created that operate fully automatically and can transmit information across long ranges.

Research in multiplex devices, which can detect multiple viruses at once, will continue to expand.

A limitation of many current devices is that sampling and analysis are usually performed as two separate steps, and the liquid sample collected from air often must be manually put into the chip for analysis. This could highly limit applicability in remote settings because a trained end user is required. To further increase automation, more microfluidic devices that directly use air for analysis (perhaps through a vacuum pump) could be developed and tested. As multiple viruses (including SARS-CoV-2 and influenza) should be monitored for public health protection while a limited number of devices can realistically be produced, spatial and spectral multiplexing capability of these devices must be improved. To create and deploy monitoring and diagnostic devices, it is important for them to be tested in the type of environments where they will ultimately be used, like in office buildings or parks. A controlled laboratory environment is not fully representative of real-world conditions, so future experiments need to emphasize indoor and outdoor field testing. The reusability J o u r n a l P r e -p r o o f and durability of devices, and reproducibility among devices of the same design, also needs to be investigated in studies, as they may be used continuously for several months or years.

COVID-19 has had profound economic and social consequences and a second pandemic could be even more devastating. Therefore, more effort must be placed on rapidly identifying novel pathogens and curbing outbreaks. Microfluidic air monitors need to become integrated with next generation sequencing (NGS) to accomplish this. Multiplexing can allow concurrent monitoring for new and existing pathogens, and public health agencies could be alerted when a novel pathogen is identified. Overall, the use of microfluidic air monitoring devices could help manage existing disease outbreaks and prevent new ones.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

COVID-19 infection: origin, transmission, and characteristics of human coronaviruses

Collection of airborne microorganisms into liquid by bubbling through porous medium

Evaluation of a novel microfluidic immunomagnetic agglutination assay method for detection of dengue virus NS1 antigen

New Sampler for the Collection, Sizing, and Enumeration of Viable Airborne Particles

Influence of Suspending Liquid, Impactor Type, and Substrate on Size-Selective Sampling of MS2 and Adenovirus Aerosols

Airborne Precautions

Rapid Airborne Influenza Virus Quantification Using an Antibody-Based Electrochemical Paper Sensor and Electrostatic Particle Concentrator

Recent progress in nanomaterial-based sensing of airborne viral and bacterial pathogens

Methods in virus diagnostics: From ELISA to next generation sequencing

Effects of air temperature and relative humidity on coronavirus survival on surfaces

Simple Strategy for Rapid and Sensitive Detection of Avian Influenza A H7N9 Virus Based on Intensity-Modulated SPR Biosensor and New Generated Antibody

SERS imaging-based aptasensor for ultrasensitive and reproducible detection of influenza virus A

Lab-on-a-chip devices for global health: Past studies and future opportunities

Development of an automated wet-cyclone system for rapid, continuous and enriched bioaerosol sampling and its application to real-time detection

Development of a flow-free magnetic actuation platform for an automated microfluidic ELISA

A digital microfluidics platform for loop-mediated isothermal amplification detection

COVID-19 Map -Johns Hopkins Coronavirus Resource Center

Aerosol transmission is an important mode of influenza A virus spread

Nanowire nanosensors for highly sensitive and selective detection of biological and chemical species. Science (80-. )

Droplet-based microfluidics detector for bioaerosol detection

Electrowetting-based digital microfluidics platform for automated enzyme-linked immunosorbent assay

The size and the duration of air-carriage of respiratory droplets and dropletnuclei

A field indoor air measurement of SARS-CoV-2 in the patient rooms of the largest hospital in Iran

The Potential Role of Smartphone-Based Microfluidic Systems Detection of COVID-19 Using Saliva Specimen

A prototype personal aerosol sampler based on electrostatic precipitation and electrowetting-on-dielectric actuation of droplets

A simple method for the detection of measles virus genome by loop-mediated isothermal amplification (LAMP)

Rapid isothermal amplification and portable detection system for SARS-CoV-2

The role of particle size in aerosolised pathogen transmission: A review

Effect of Impaction, Bounce and Reaerosolization on the Collection Efficiency of Impingers

Coronavirus (COVID-19): A Review of Clinical Features

Drivers of airborne human-to-human pathogen transmission

A Method for Particle Size Amplification by Water Condensation in a Laminar

Moderated, Water-Based, Condensational Particle Growth in a Laminar Flow

A Laminar-Flow, Water-Based Condensation Particle Counter (WCPC)

ELISA: Methods and Protocols

Size distribution of exhaled particles in the range from 0.01 to 2.0μm

Continuous aerosol size separator using inertial microfluidics and its application to airborne bacteria and viruses

One-step rapid quantification of SARS-CoV-2 virus particles via low-cost nanoplasmonic sensors in generic microplate reader J o u r n a l P r e -p r o o f and point-of-care device

Diagnostic value of respiratory virus detection in symptomatic children using real-time PCR

Detection of COVID-19: A review of the current literature and future perspectives

Microfluidic Device for Efficient Airborne Bacteria Capture and Enrichment

Hospital indoor air quality monitoring for the detection of SARS-CoV-2 (COVID-19) virus

Routes of influenza transmission

Highly sensitive detection of influenza virus with SERS aptasensor

Rapid and Sensitive Detection of H1N1/2009 Virus from Aerosol Samples with a Microfluidic Immunosensor

Bioaerosol sampling of a ventilated patient with COVID-19

Comparing the performance of 3 bioaerosol samplers for influenza virus

Microfluidic System for Rapid Detection of Airborne Pathogenic Fungal Spores

Development of a SERS-based lateral flow immunoassay for rapid and ultra-sensitive detection of anti-SARS-CoV-2 IgM/IgG in clinical samples

A semi-quantitative method for point-of-care assessments of specific pathogenic bioaerosols using a portable microfluidics-based device

A novel microfluidic module for rapid detection of airborne and waterborne pathogens

First airborne pathogen direct analysis system

Real-system

Catalytically Active Single-Chain Polymeric Nanoparticles: Exploring Their Functions in Complex Biological Media

Reciprocatingflowing on-a-chip enables ultra-fast immunobinding for multiplexed rapid ELISA detection of SARS-CoV-2 antibody

Viruses: Structure, Function and Uses

Essential Human Virology

Genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding

COVID-19 patients in earlier stages exhaled millions of SARS-CoV-2 per hour

Microfluidic systems for pathogen sensing: A review

Detection of severe acute respiratory syndrome coronavirus 2 and influenza viruses based on CRISPR-Cas12a

Integrated quantum dot barcode smartphone optical device for wireless multiplexed diagnosis of infected patients

A microfluidics-based on-chip impinger for airborne particle collection

Size distribution and sites of origin of droplets expelled from the human respiratory tract during expiratory activities

Environmental contamination of SARS-CoV-2 on surfaces, air-conditioner and ventilation systems

r n a l P r e -p r o o f respirators-surgical-masks-and-face-masks

Point-of-care microfluidic devices for pathogen detection

Detection of bacterial and viral pathogens using photonic point-of-care devices

Susceptibility of an airborne common cold virus to relative humidity

Assessment of SARS-CoV-2 serological tests for the diagnosis of COVID-19 through the evaluation of three immunoassays: Two automated immunoassays (Euroimmun and Abbott) and one rapid lateral flow immunoassay (NG Biotech)

Estimation of the asymptomatic ratio of novel coronavirus infections (COVID-19)

Loop-mediated isothermal amplification of DNA

Design and performance low-cost micro-channel aerosol collector

Evolution in Lateral Flow-Based Immunoassay Systems

Polymerase Chain Reaction Assay for Detection of Human Cytomegalovirus Downloaded from

Polymerase chain reaction assay for detection of human cytomegalovirus

A water cyclone to preserve insoluble aerosols in liquid flow -An interface to flow cytometry to detect airborne nucleic acid

Collection, particle sizing and detection of airborne viruses

Aerosol sampling using an electrostatic precipitator integrated with a microfluidic interface

Continuous and real-time bioaerosol monitoring by combined aerosol-to-hydrosol sampling and ATP bioluminescence assay

Field-Effect Sensors Detection: From Ebola to SARS-CoV-2 and Plant Viral Enhancers

Multiplex, quantitative, reverse transcription PCR detection of influenza viruses using droplet microfluidic technology

Rapid and Fully Microfluidic Ebola Virus Detection with CRISPR-Cas13a

Computational sensing of herpes simplex virus using a cost-effective on-chip microscope

Holographic detection of nanoparticles using acoustically actuated nanolenses

SARS-CoV-2 RNA detection in the air and on surfaces in the COVID-19 ward of a hospital in Milan

Characterization of reaerosolization from impingers in an effort to improve airborne virus sampling

Handheld Point-of-Care System for Rapid Detection of

Extracted RNA in under 20 min

Smartphone-based fluorescent lateral flow immunoassay platform for highly sensitive point-of-care detection of Zika virus nonstructural protein 1

Electrohydrodynamic enhanced transport and trapping of airborne particles to a microfluidic air-liquid interface

Introduction to Surface Plasmon Resonance

Causative Virus (SARS-CoV-2) in Human Nasopharyngeal Swab Specimens Using Field-Effect Transistor-Based Biosensor

Correction: Rapid Detection of COVID-19 Causative Virus (SARS-CoV-2) in Human Nasopharyngeal Swab Specimens Using Field-Effect Transistor-Based Biosensor

Integrating silicon nanowire field effect transistor, microfluidics and air sampling techniques for real-time monitoring biological aerosols

Rapid flu diagnosis using silicon nanowire sensor

A critical review on PCR, its types and applications

Super-spreaders and the rate of transmission of the SARS virus. Phys. D Nonlinear Phenom

Microfluidics: Fundamentals, Devices, and Applications

A Temperature-Controlled AGI-30 Impinger for Sampling of Bioaerosols

Smartphone-based multiplex 30-minute nucleic acid test of live virus from nasal swab extract

Development of an automated electrostatic sampler (AES) for bioaerosol detection

A fully automated centrifugal microfluidic system for sample-to-answer viral nucleic acid testing

Ultrasensitive electrical detection of hemagglutinin for point-of-care detection of influenza virus based on a CMP-NANA probe and top-down processed silicon nanowire field-effect transistors

Surface plasmon resonance-based real-time bioaerosol detection

How Seeing Became Knowing: The Role of the Electron Microscope in Shaping the Modern Definition of Viruses

Methods for Sampling of Airborne Viruses

Magnetic SERS Strip for Sensitive and Simultaneous Detection of Respiratory Viruses

Rapid detection of multiple respiratory viruses based on microfluidic isothermal amplification and a real-time colorimetric method

Fluorescent Imaging of Single Nanoparticles and Viruses on a Smart Phone

Airborne transmission of severe acute respiratory syndrome coronavirus-2 to healthcare workers: a narrative review

Four point-of-care lateral flow immunoassays for diagnosis of COVID-19 and for assessing dynamics of antibody responses to SARS-CoV-2

A novel double antibody sandwich-lateral flow immunoassay for the rapid and simple detection of hepatitis C virus

Exhaled droplets due to talking and coughing

Rapid Differential Diagnosis of Seven Human Respiratory Coronaviruses Based on Centrifugal Microfluidic Nucleic Acid Assay

Kinetic ELISA in microfluidic channels

Tunable and label-free virus enrichment for ultrasensitive virus detection using carbon nanotube arrays

Minimum Infective Dose of the Major Human Respiratory and Enteric Viruses Transmitted Through Food and the Environment

Chemiluminescence immunoassay

Advanced "lab-on-a-chip" to detect viruses -Current challenges and future perspectives

Investigations on DNA intercalation and surface binding by SYBR Green I, its structure determination and methodological implications

The authors would like to thank the Natural Sciences and Engineering Research Council of Canada (NSERC) for financial support provided to PR.